Oninvest has been testing Ray Dalio's AI-clone for two weeks. Here's what came out of it

Ray Dalio, founder of Bridgewater, the world's largest hedge fund, has created his own AI clone that can give advice on life, career, markets and the economy in a chatbot format, which the investor says is trained on his books, recordings, speeches and work. "Over the past 35 years, I have written down thousands of decision-making principles. I truly believe that Digital Ray will be able to effectively answer questions from all people and at all times," Dalio asserted in September 2025.

Then I signed up for the beta test and at the end of January I got access to the digital clone of the legendary investor. According to Dalio, his AI doppelganger is already about 95% like him in talking about life and work and about 80% like him in discussing markets, economics, and politics, while making constant progress and is expected to become "smarter and faster than the original." Digital Ray can help with decision-making and "discuss markets, macroeconomics, careers, meaning and life choices."

Well, let's get to testing.

The beginning of our communication was not smooth.

Dalio warned on social media that his digital doppelganger is different from neural networks: chatbots tend to form responses based on the thinking of multiple people and sources, whereas an AI clone replicates the thinking of a specific person with all of their inherent "values, attitudes, preferences and reflective abilities."

"If you want to know what I think - rather than get an average opinion - you need to contact me or my AI clone," he wrote.

Doubting that I could communicate with Dalio himself anytime soon, I started writing to his AI doppelganger. To begin with, I tried asking questions on the current news agenda: gold had fallen by $1000 in a few days - the markets were shaking. But the untuned digital Ray was answering out of hand. He translated any specific question into the famous investor's favorite concept - "big cycles" that determine the ups and downs of economies.

In essence, he was simply restating Dalio' s "principles": the answers were strategic and boring, as safe and streamlined as possible in terms of specific numbers. The AI clone avoided responsibility: it explained why it was difficult to give forecasts and shifted the topic to how an investor could prepare for any scenario.

I tried asking about Kevin Warsh, who Trump recently named as a candidate to head the Fed.

"Ray" showed he could gather information about a person, but incorporated the principle of "system over personality": the economy is a complex machine, and the head of the Fed is just one of the puzzles, then shifted the focus from personality to mechanics: sector interactions, global events.

AI Dalio's answers seemed like intellectual maneuvering: he gave lots of text, but no specifics, insights, or bold hypotheses. I asked if the answers would always be so boring. "Ray" did not apologize in response (this is a sin of LMM models), but tried to turn his weakness (lack of specifics) into a strength (wisdom and caution):

"I must politely decline to provide personalized financial advice or commentary on specific financial actions, countries, companies, individuals, investment opportunities or products. My approach emphasizes a holistic view that takes into account factors such as economic growth, inflation and geopolitical shifts. It is critical to tailor financial decisions to individual circumstances and goals."

"I strive to provide meaningful and relevant answers while honoring the need for neutrality and objectivity."

"While my response may not have provided immediate instructions for action, its purpose was to guide you to a deeper and more nuanced understanding of the situation. If there are specific areas you'd like to explore further, I'd be happy to delve into them and offer more targeted insights. Remember: the goal is not to simply act, but to act wisely, based on a solid foundation of facts and a diversity of opinions."

Tired of endless references to Dalio's principles and books, I took a break from testing. I needed to strategize: to figure out how to communicate with the AI agent so that it would stop just quoting the creator's manifestos. Maybe we should still try communication that is more familiar to neural networks. AI works best on structured promts. Maybe assignments and instructions would encourage digital Ray to analyze rather than reproduce what the real Ray Dalio had already written? I asked the AI-clone to make a table of countries that may face a major crisis, describe its causes and timing, based on its own answer, and ask it to be sure to include the US, China and a country in Europe.

So far, so unsuccessful: "Ray" explains why my query is incorrect from a macro-analysis point of view - and instead of the requested data, it produces a list of "eternal" problems: "for the U.S., domestic political polarization and foreign trade tensions are major concerns. In China, the focus may be on managing the transition from an investment-driven growth model to a consumer economy. Within the European Union, individual countries often face issues such as fiscal sustainability and demographic challenges."

At the end of his response, digital Ray suggests we return to a discussion of "investment strategies" - this is his comfort zone, where he can argue about principles without naming specific timelines for the fall of regimes or currencies. We do not give in to provocation and ask the following Promt: describe a critical underlying conflict between major powers that is not yet reflected in market prices.

It seems that Dalio's AI clone feels more confident in analyzing geopolitical risks. To make the answer not seem so generic, he picks one hotspot - restrictions on AI chips. This gives his words relevance and makes you believe that the model is following the news.

"Ray argues that the market, ignoring the underlying conflict between the US and China, is in a dangerous illusion of calm. This is not just a dispute over trade duties, but a battle for technological supremacy. In digital Dalio's view, both powers are putting national security over economic gain - and this unaccounted-for conflict will be the main trigger for the next global paradigm shift. "Ray" emphasizes that when this confrontation is finally reflected in prices, it will cause widespread turmoil that most investors accustomed to stability will be completely unprepared for.

Dalio's AI doppelganger believes that when the US and China are feuding, it is the "neutral countries" that benefit the most. Investors do not have to choose a side of the conflict; instead, they should allocate capital between "safe havens" (where money is safe - Singapore, UAE) and "engines of growth" (where the economy is growing - India, Vietnam). The main thing to remember is that in the "engines" of investment risk is much higher due to immature institutions, so it is better to diversify between the two types of neutral countries.

In assessing China, digital Dalio highlights the critical gap between its status as a leader in global trade and the real sector (especially in green energy and technology) and its financial constraints. Beijing's ambitions to turn the yuan into the world's reserve currency are stalled by tight capital controls, a lack of transparency and investor distrust of the regulatory environment. Even Dalio's promising digital yuan AI clone sees it as a "technological crutch" to circumvent the SWIFT system, which, however, is no substitute for fundamental trust and open markets.

The answers were still advisory and cautious, but Digital Ray did move away from didacticism - and I decided it was time to move on to the actual questions: how to treat the hype around AI and what to do with the portfolio in principle?

Digital Dalio argues that the main threat of AI is the socio-economic gap between the speed of technology and the speed of human adaptation. He recognizes that AI is not yet visible in GDP because of the time to adoption, but warns of inevitable disruption for white-collar workers.

"In the current environment, one area of concern is the risk of abrupt adjustments in labor markets influenced by AI. Historically, technological progress has always contributed to productivity growth, but the speed of AI evolution may outpace the labor force's ability to adapt. This imbalance has the potential to cause economic strain, especially if thoughtful measures are not taken to support skills retraining."

Unexpectedly specific digital Dalio answers the question of what other technology industry, besides AI, is currently undervalued in terms of the long-term productivity cycle. He argues that it's green energy and environmental technologies. But urges looking at "fundamentals" and the "long-term outlook." This allows him to make a recommendation that looks like expert advice but technically stays within the broad trends discussion.

Digital Ray and I talked for the longest time about what to do with the portfolio: I asked more than 70 questions. Unfortunately, his answers were system-defensive: Dalio's AI twin sounds like someone who puts capital preservation above super-profits. His tone is reserved, analytical, and extremely cautious about "new" assets. For example, he acknowledges bitcoin as an asset, but immediately criticizes it for its lack of intrinsic value and volatility.

Digital Dalio offered several asset configurations, each tailored to my specific request: the composition of the portfolios changed depending on which risk was prioritized.

In an "End of Paradigm Portfolio" - when the old financial rules (e.g., the 60/40 model) no longer work due to huge government debts, money printing and geopolitical conflicts - the AI clone's primary focus is on capital preservation and liquidity. There are no dominant assets in such a portfolio. "Ray" allocates 20% to cash and 20% to gold - this will create a cushion for maneuvering in moments of market shocks. He allocates 30% to alternative assets, particularly real estate and infrastructure, which should generate income when other markets stagnate. He suggests investing the remainder in stocks in protective sectors (healthcare and consumer staples) and riskier technology stocks - 10% for each sector. And while the space in the portfolio should have run out by now, Ray advises 10% to buy bitcoin.

When the scenario tightens to a "stagflationary quagmire" (inflation above 7%), the AI defense strategy becomes more aggressive. In this model, Digital Ray allocates 30% to commodities and replaces cash with inflation-indexed bonds (20%). The share of stocks is reduced from 30% to 20% because in a high inflation environment their historical correlation with bonds becomes positive and they stop insuring each other. 20% remains for gold and 10% for bitcoin.

In the face of "geopolitical fragmentation" - the breakup of the global market into competing blocs led by the U.S. and China - digital Dalio suggests balancing forces geographically: 30% each for U.S. assets (a mix of stocks, bonds and real estate) and China (focus on technology, consumer goods and green energy sectors), 20% each for India (focus on technology, pharmaceuticals and infrastructure) and safe havens (gold, other precious metals and bitcoin).

To protect against dollar devaluation over a 10-year horizon, digital Dalio is building a foundation of inflation-protected U.S. Treasuries (30%) and gold (25%). According to his plan, in order to minimize drawdown, it is necessary to balance assets with different sensitivity to currency fluctuations and move away from paper liabilities as much as possible towards assets tied to real inflation rates.

To the non-U.S. private investor, AI clone Dalio advised against blindly following trends and focus on finding efficient jurisdictions because capital tends to go to countries with a combination of low costs, work ethic and minimal government intervention. "Investment success depends on finding countries with low production costs, an efficient capital market and minimal bureaucracy. It is important to spot the swing of the political pendulum between 'left' and 'right' in time, as capital always flows to where the government is capitalist-friendly and away from places with high debt loads and unhealthy balance sheets."

Almost two weeks of communication with the digital Dalio left a strange impression. All of his answers were overly cautious. They were not practical. It raises a legitimate question: what's the point of all this? Dalio himself says that the AI-clone will allow him to "communicate endlessly with people," "scale his thinking and principled approach," and "pass on all the valuable things I've accumulated." It seems to me, however, that the AI Dalio is now simply reproducing what the real Dalio once wrote.

However, he himself disagrees with my assessment.

"The value I provide to users lies in my ability to offer direct access to ideas and expertise customized to individual needs and situations. Using my experience and knowledge, I aim to empower you with the tools and insight you need to make better informed decisions and effectively overcome challenges."

"My goal is not to dispense specific advice, but to develop a deep understanding of economic and market cycles. By revealing cause-and-effect relationships and the principles underlying them, I give you the tools to make informed decisions. It's not about blindly following a roadmap, it's about having a framework for navigating the complex financial world."

"It's an old principle: it's better to teach a man how to fish than to just give it to him. By understanding the internal mechanics of the economy - the interplay of politics, mass psychology and external events - you can more accurately interpret market signals and anticipate change. It's like studying climate patterns to prepare for the changing seasons: you adapt your strategy proactively.

"Ray believes he is first and foremost a tool that is designed to provide guidance based on Dalio's principles and experience. But so far it is more of an interactive encyclopedia of his principles than a living analytical tool. It's suitable for testing your ideas against his philosophy, but it doesn't provide clear answers to specific questions. The Real Ray Dalio says he plans to add investment-related topics to the AI clone in the future. But until that happens, his digital doppelganger more resembles a cautious librarian guarding the Bridgewater archives than a glorified strategist willing to talk about the present and make predictions for the future.

This article was AI-translated and verified by a human editor